A real-time monitoring method and system for food storage environment

Abstract

The present invention relates to the field of food safety technology, specifically to a real-time monitoring method and system for a food storage environment, comprising the following steps: based on real-time monitoring data of the food storage environment, obtaining temperature, humidity, and light parameter state values, collecting multi-parameter states, comparing them with storage standards respectively, adjusting the fan operating speed and the opening time of the refrigeration equipment according to the comparison difference values, and generating dynamic adjustment results of environmental parameters. In the present invention, by obtaining environmental parameters such as temperature, humidity, and light, and comparing them with standard storage conditions, the operation of the fan and refrigeration equipment is adjusted in real time and dynamically, the storage environment conditions are optimized, and the food is ensured to be kept in an ideal state. A neural network is used to map the environmental parameters to the food shelf life state, and the parameter settings are optimized step by step. The impact of environmental quality on food is accurately predicted to ensure the best matching of storage conditions, and the optimal configuration is selected by cross-combination scoring, so that the equipment response is more sensitive and energy waste is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of food safety, and in particular to a real-time monitoring method and system for a food storage environment. Background Art

[0002] The field of food safety technology involves food quality and safety control technologies from production, processing, storage to consumption. This field uses a series of scientific methods and technical measures to prevent food from being contaminated or spoiled at any stage to ensure that food is harmless to human health. The main technologies include microbial detection, chemical contaminant detection, physical safety detection, and the establishment of traceability systems. In addition, food safety technology also includes the development of new preservation technologies and improved packaging materials to extend the shelf life of food and maintain the stability of food nutrients. With the advancement of technology and the improvement of consumer health awareness, the field of food safety continues to integrate new technologies, such as the Internet of Things, big data, and artificial intelligence, to improve the efficiency and accuracy of food safety management.

[0003] Among them, the real-time monitoring method of food storage environment refers to the use of sensors and other measuring devices to monitor key parameters in the food storage environment in real time, such as temperature, humidity, gas composition, etc., to ensure that these conditions are always in an ideal state to prevent food spoilage or contamination. The main purpose of this monitoring system is to improve the level of food safety, reduce food losses caused by improper storage conditions, and provide a scientific basis for food safety risk management through continuous data collection. Such systems are widely used in food warehouses, cold chain logistics, supermarkets and any large-scale food storage and transportation sites, and are of great significance to protecting public health and quality control in the food industry.

[0004] Existing technologies are mostly based on independent parameter collection based on environmental monitoring, lacking the mapping relationship between the environment and the shelf life of food, and relying only on over-standard alarms, unable to make real-time dynamic adjustments. The delayed response makes it impossible to respond quickly when the temperature and humidity fluctuate drastically, resulting in delayed environmental adjustment and affected food quality. In addition, the existing system cannot accurately predict and adjust storage environment changes by accumulating historical data, lacks intelligent management of storage conditions, and increases the risk of food loss. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a real-time monitoring method and system for food storage environment.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a method for real-time monitoring of food storage environment, comprising the following steps:

[0007] S1: Based on the real-time monitoring data of the food storage environment, the temperature, humidity, and light parameter status values ​​are obtained, and the multi-parameter status is collected and compared with the storage standards respectively. The fan speed and the opening time of the refrigeration equipment are adjusted according to the comparison difference value to generate the dynamic adjustment result of the environmental parameters;

[0008] S2: Based on the dynamic adjustment result of the environmental parameters, the environmental parameters and the food shelf life status are mapped, the environmental parameters are input into the neural network connection structure, the weights and deviations are adjusted step by step, the node parameters are updated, and the environmental quality impact prediction value is obtained;

[0009] S3: Based on the predicted value of environmental quality impact, the number of layers, number of neurons and learning rate of the neural network are cross-combined, the matching degree of each combination is scored, the combination structure with the highest matching degree is selected, and the optimal neural network configuration result is established;

[0010] S4: Based on the optimal neural network configuration result, input environmental parameters to the neural network in real time, predict the impact of environmental parameters on food quality one by one, readjust the input parameters, use the adjusted parameters to calculate the output power and duration of the equipment, and generate environmental parameter correction values;

[0011] S5: Based on the environmental parameter correction value, call the equipment operation status and environmental feedback value, monitor a new round of environmental parameters, combine temperature, humidity, light and equipment control status, accumulate all equipment output and monitoring data, and generate food storage risk assessment indicators.

[0012] The dynamic adjustment results of the environmental parameters include the fan operating speed, the refrigeration equipment startup time, and the environmental parameter difference value; the environmental quality impact prediction value includes the temperature prediction value, the humidity prediction value, and the light prediction value; the optimal neural network configuration result includes the number of neural network layers, the number of neurons, and the learning rate; the environmental parameter correction values ​​include the equipment output power and the equipment output time; the food storage risk assessment indicators include temperature risk, humidity risk, light risk, and equipment operating status.

[0013] As a further solution of the present invention, the step of obtaining the result of dynamic adjustment of the environmental parameters is specifically as follows:

[0014] S111: collect real-time monitoring data, including temperature, humidity, and light parameters, obtain the current state value of the food storage environment through the sensor network, and generate real-time environmental status records;

[0015] S112: Compare the temperature, humidity, and light conditions in the real-time environmental status record with the preset storage standard, calculate the difference values ​​of multiple parameters, and obtain the environmental difference value result;

[0016] S113: According to the environmental difference value result, the fan speed and the working time of the refrigeration equipment are adjusted using the formula:

[0017] ;

[0018] Calculate the adjustment amount of fans and refrigeration equipment, and generate dynamic adjustment results of environmental parameters ;

[0019] in, represents the temperature difference value, Represents the humidity difference value, Represents the illumination difference value, is the standard threshold value of temperature, is the standard threshold value of humidity, is the standard threshold of illumination, is the temperature regulation coefficient, is the humidity adjustment factor, is the light regulation coefficient.

[0020] As a further solution of the present invention, the step of obtaining the environmental quality impact prediction value is specifically as follows:

[0021] S211: extracting key temperature, humidity, and light parameters based on the dynamic adjustment result of the environmental parameters, inputting the parameters into the input layer of the neural network through the data interface, and generating an input layer data carrier;

[0022] S212: Transmitting the data layer by layer to each hidden layer of the neural network through the input layer data carrier, adjusting the weight and bias of the nodes of each layer in real time, and using the adjusted weight and bias of each layer to generate intermediate node parameters;

[0023] S213: Based on the intermediate node parameters, the formula is adopted:

[0024] ;

[0025] Adjust weights and biases, update the weight of each node, and obtain updated node parameters;

[0026] in, Represents the weight before adjustment, which is the main parameter for transmitting information in the neural network and directly affects the output of the network. Represents the adjusted weight, which is the new weight value obtained by adjusting the algorithm update to optimize the prediction or classification accuracy of the network. Represents the learning rate, controls the step size of weight adjustment, and directly affects the speed and stability of network training. It is the derivative of the loss function with respect to the weight, indicating the sensitivity of the loss function to the change under the current weight, and is used to guide the optimization direction of the weight. is a regularization parameter used for weight attenuation, preventing network overfitting and enhancing the generalization ability of the model. is the weight decay coefficient, which is used to adjust the influence of each weight and reduce the problem of network dependence on specific paths caused by excessive weights;

[0027] S214: Using the updated node parameters and combining with the data processing of the network output layer, the influence of the environmental parameters on the shelf life of the food is calculated, and the environmental quality impact prediction value is output.

[0028] As a further solution of the present invention, the step of obtaining the optimal neural network configuration result is specifically:

[0029] S311: Based on the environmental quality impact prediction value, extract the number of layers, number of neurons and learning rate of each network configuration through the system interface, create differentiated network architecture combinations, and integrate them into a network architecture combination list;

[0030] S312: Perform independent network training on each configuration in the network architecture combination list, calculate the performance error of each network configuration using the validation set, evaluate the effectiveness of each network configuration through the performance error, and generate effectiveness evaluation data;

[0031] S313: Analyze the performance evaluation data and make a comprehensive score based on the performance error and network complexity of each configuration, using the formula:

[0032] ;

[0033] Based on the scoring results, the network configuration with the highest matching degree is selected to obtain the optimal configuration result;

[0034] in, Indicates The performance error of each configuration is Represents the complexity of the configuration, is the adjustment coefficient, which is used to balance the influence of performance error and network complexity in the total score. Represents the calculated comprehensive score, which comprehensively considers the performance and complexity of all network configurations and is used to select the optimal neural network configuration. Represents the total number of network configurations involved in the scoring calculation, reflecting the number of evaluated network architecture combinations;

[0035] S314: Using the optimal configuration result, the neural network architecture is adjusted to ensure the network configuration is optimized and to establish the optimal neural network configuration result.

[0036] As a further solution of the present invention, the step of obtaining the modified device control output is specifically:

[0037] S411: Based on the optimal neural network configuration result, input environmental parameters to the neural network structure in real time, perform food quality impact prediction for the input temperature, humidity, and light environmental parameters, and generate food quality prediction results;

[0038] S412: Using the food quality prediction result, compare the relationship between the input environmental parameters and the predicted impact, calculate the parameter adjustment amount, readjust the environmental parameters, and calculate the output power and operating time of the equipment, using the formula:

[0039] ;

[0040] Obtaining environmental parameter correction values;

[0041] in, Represents the power adjustment factor, which adjusts the power output based on the combined effect of all environmental parameters. Represents the difference between the adjusted temperature parameter and the standard temperature, Represents the difference between the adjusted humidity parameter and the standard humidity, Represents the difference between the adjusted lighting parameters and the standard lighting. Represents the maximum allowable value of temperature, Represents the maximum allowable value of humidity, Represents the weight parameter of the lighting adjustment effect, and adjusts the device power requirement according to the weight of the lighting parameter. Represents the power requirement of the device.

[0042] As a further solution of the present invention, the steps for obtaining the food storage risk assessment index are specifically as follows:

[0043] S511: Based on the corrected device control output, the device operation status and environmental feedback data are acquired in real time, and the temperature, humidity, and light parameters are collected, combined with the device control status, to form an environmental monitoring data packet;

[0044] S512: Analyze the environmental monitoring data packet, use data fusion technology to integrate the real-time environmental parameters with the equipment output, calculate the deviation by comparing the preset threshold value with the actual value, and evaluate its potential impact on food quality to obtain preliminary risk assessment data;

[0045] S513: Analyze the preliminary risk assessment data using the formula:

[0046] ;

[0047] Calculate comprehensive risk indicators, quantify food storage risks, and generate food storage risk assessment indicators;

[0048] in, Represents the actual environmental parameters being monitored, including temperature, humidity, and light. Represents the corresponding safety standard parameters, which are preset environmental parameter standards used to ensure the safety of the food storage environment. is the risk weight of multiple parameters, Represents the food storage risk assessment index, which is a comprehensive risk value. Represents the total number of environmental parameters.

[0049] A food storage environment real-time monitoring system, the food storage environment real-time monitoring system is used to execute the above-mentioned food storage environment real-time monitoring method, the system comprises:

[0050] The dynamic adjustment module of environmental parameters obtains the temperature, humidity and light status values ​​based on the real-time monitoring data of the food storage environment, compares the parameters with the standard values, adjusts the fan speed and the duration of the refrigeration equipment, dynamically adjusts the environmental combination, and generates the dynamic adjustment results of environmental parameters;

[0051] The neural network configuration optimization module dynamically adjusts the environmental parameters based on the environmental parameters, associates and maps the environmental parameters with the food shelf life status, inputs the environmental parameters into the neural network for weight and bias adjustment, calculates the impact of the environment on the food shelf life, and obtains the optimal neural network configuration result;

[0052] The environmental parameter correction and quality prediction module inputs the real-time monitored environmental parameters based on the optimal neural network configuration result, calculates the impact on food quality, corrects the environmental parameters, recalculates the equipment output power and duration, and generates environmental parameter correction values;

[0053] The food storage risk assessment module monitors the equipment operation status, obtains the environmental parameter feedback value, and generates food storage risk assessment indicators based on the environmental parameter correction value, the accumulated fan and refrigeration equipment output data combined with the equipment control status.

[0054] Compared with the prior art, the advantages and positive effects of the present invention are:

[0055] In the present invention, by obtaining environmental parameters such as temperature, humidity, and light, and comparing them with standard storage conditions, the operation of fans and refrigeration equipment is adjusted dynamically in real time to optimize storage environmental conditions and ensure that food remains in an ideal state. A neural network is used to map environmental parameters to the shelf life of food, and the parameter settings are optimized step by step to accurately predict the impact of environmental quality on food and ensure the best match of storage conditions. Cross-combination scoring selects the optimal configuration, making the equipment more responsive and reducing energy waste. Closed-loop monitoring combines equipment output and environmental feedback, accumulates data to generate risk assessments, quantifies storage risks, and improves food safety and management efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0057] Figure 2 A flowchart of the steps for obtaining the results of dynamic adjustment of environmental parameters of the present invention;

[0058] Figure 3 A flow chart of the steps for obtaining the predicted value of environmental quality impact of the present invention;

[0059] Figure 4 A flowchart of the steps for obtaining the optimal neural network configuration result of the present invention;

[0060] Figure 5 A flowchart of the steps for obtaining the device control output after the modification of the present invention;

[0061] Figure 6 The present invention is a flow chart of the steps for obtaining food storage risk assessment indicators. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0063] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0064] Example 1: Please refer to Figure 1 The present invention provides a technical solution: a method for real-time monitoring of a food storage environment, comprising the following steps:

[0065] S1: Based on the real-time monitoring data of the food storage environment, the temperature, humidity, and light parameter status values ​​are obtained, and the multi-parameter status is collected and compared with the storage standards respectively. The fan speed and the opening time of the refrigeration equipment are adjusted according to the comparison difference value to generate the dynamic adjustment result of the environmental parameters;

[0066] S2: Based on the results of dynamic adjustment of environmental parameters, the environmental parameters and food shelf life are mapped, the environmental parameters are input into the neural network connection structure, the weights and biases are adjusted step by step, the node parameters are updated, and the predicted value of environmental quality impact is obtained;

[0067] S3: Based on the predicted value of environmental quality impact, the number of layers, number of neurons and learning rate of the neural network are cross-combined, the matching degree of each combination is scored, the combination structure with the highest matching degree is selected, and the optimal neural network configuration result is established;

[0068] S4: Based on the optimal neural network configuration result, input environmental parameters into the neural network in real time, predict the impact of environmental parameters on food quality one by one, readjust the input parameters, use the adjusted parameters to calculate the output power and duration of the equipment, and generate environmental parameter correction values;

[0069] S5: Based on the environmental parameter correction values, call the equipment operation status and environmental feedback values, monitor the new round of environmental parameters, combine temperature, humidity, light with equipment control status, accumulate all equipment output and monitoring data, and generate food storage risk assessment indicators.

[0070] The dynamic adjustment results of environmental parameters include fan operating speed, refrigeration equipment startup time, and environmental parameter difference values; the environmental quality impact prediction values ​​include temperature prediction value, humidity prediction value, and light prediction value; the optimal neural network configuration results include the number of neural network layers, the number of neurons, and the learning rate; the environmental parameter correction values ​​include equipment output power and equipment output time; and the food storage risk assessment indicators include temperature risk, humidity risk, light risk, and equipment operating status.

[0071] See also Figure 2 , the specific steps for obtaining the results of dynamic adjustment of environmental parameters are:

[0072] S111: collect real-time monitoring data, including temperature, humidity, and light parameters, obtain the current state value of the food storage environment through the sensor network, and generate real-time environmental status records;

[0073] The real-time monitoring system continuously collects temperature, humidity and light intensity data in the food storage environment. These data are uploaded to the central monitoring system in real time by multiple sensors distributed in the storage area, and the status records are updated in real time through the data interface. Each sensor has a self-calibration function to ensure the accuracy and real-time nature of the data. The frequency of data collection is high, up to once per minute, to ensure that the status records obtained can truly reflect the immediate changes in the environment. These real-time data will be stored in the database for subsequent processing and provide decision support for the environmental management system. These records are updated in real time to ensure that the system can quickly respond to any slight changes in the environment.

[0074] S112: Compare the temperature, humidity, and light conditions in the real-time environmental status record with the preset storage standard, calculate the difference values ​​of multiple parameters, and obtain the environmental difference value result;

[0075] The real-time environmental status records are compared with the preset standards of the storage environment, mainly comparing the three key indicators of temperature, humidity and light. In this process, the system reads the threshold data in the environmental standard setting database and compares it with the real-time collected data in real time. The difference calculation uses a simple subtraction operation, that is, the difference between the measured value and the standard value. In addition, the system will classify the comparison results to determine which differences exceed the allowable range. These classification results will be recorded in the environmental difference value report to provide a basis for the adjustment of refrigeration equipment and fans. This method ensures the timeliness of processing and the accuracy of difference response.

[0076] S113: According to the environmental difference value result, the fan speed and the working time of the refrigeration equipment are adjusted using the formula:

[0077] ;

[0078] Calculate the adjustment amount of fans and refrigeration equipment, and generate dynamic adjustment results of environmental parameters ;

[0079] in, represents the temperature difference value, Represents the humidity difference value, Represents the illumination difference value, is the standard threshold value of temperature, is the standard threshold value of humidity, is the standard threshold of illumination, is the temperature regulation coefficient, is the humidity adjustment factor, is the light regulation coefficient.

[0080] formula:

[0081] ;

[0082] The benefit of the formula is that it combines the real-time monitoring data of three key environmental variables, temperature, humidity and light, with the preset standard differences. By taking a weighted average approach, it accurately adjusts the operating parameters of fans and refrigeration equipment, optimizes energy efficiency and maintains the stability of the storage environment.

[0083] Detailed explanation of the formula and the process of formula calculation and derivation:

[0084] Assume that the currently monitored temperature is 20°C, the standard temperature is 25°C, the humidity monitoring value is 50%, the standard humidity is 40%, the light intensity is 3000lux, and the standard light intensity is 2800lux. The adjustment coefficients are , , .

[0085] First calculate the temperature difference ratio:

[0086] ;

[0087] Humidity Difference Ratio:

[0088] ;

[0089] Light Difference Ratio:

[0090] ;

[0091] Then substitute these values ​​into the formula to calculate:

[0092] ;

[0093] The results show that according to the current environmental monitoring data, the adjustment amount of fans and refrigeration equipment should be increased by 0.0255 units to meet the environmental standards, which shows the flexibility of the system in responding to environmental changes and the ability of precise control.

[0094] See also Figure 3 , the specific steps for obtaining the predicted value of environmental quality impact are:

[0095] S211: extracting key temperature, humidity, and light parameters based on the dynamic adjustment results of environmental parameters, inputting the parameters into the input layer of the neural network through the data interface, and generating an input layer data carrier;

[0096] Based on the dynamic adjustment results of environmental parameters, temperature, humidity and light data are collected in real time by precise monitoring equipment and sent to the system through a dedicated data transmission interface. The input layer processes this data and standardizes it to match the requirements of the network training model. The adjusted data is the input of the neural network. This process ensures the real-time update and accuracy of the data, lays the foundation for subsequent deep learning training, and generates the input layer data carrier.

[0097] S212: Transmitting the data layer by layer to each hidden layer of the neural network through the input layer data carrier, adjusting the weight and bias of the nodes of each layer in real time, and using the adjusted weight and bias of each layer to generate intermediate node parameters;

[0098] The input layer data carrier contains the environmental parameters that have been preliminarily processed. These parameter data are transmitted layer by layer through the structure of the neural network. The nodes of each layer adjust the weights and biases according to the input data. This layer-by-layer transmission and adjustment realizes the deep feature extraction of the data. The output of each layer serves as the input of the next layer. Through iterative optimization, each layer finally generates adjusted intermediate node parameters to provide support for the output layer decision.

[0099] S213: Based on the intermediate node parameters, the formula is:

[0100] ;

[0101] Adjust weights and biases, update the weight of each node, and obtain updated node parameters;

[0102] in, Represents the weight before adjustment, which is the main parameter for transmitting information in the neural network and directly affects the output of the network. Represents the adjusted weight, which is the new weight value obtained by adjusting the algorithm update to optimize the prediction or classification accuracy of the network. Represents the learning rate, controls the step size of weight adjustment, and directly affects the speed and stability of network training. It is the derivative of the loss function with respect to the weight, indicating the sensitivity of the loss function to the change under the current weight, and is used to guide the optimization direction of the weight. is a regularization parameter used for weight attenuation, preventing network overfitting and enhancing the generalization ability of the model. is the weight decay coefficient, which is used to adjust the influence of each weight and reduce the problem of network dependence on specific paths caused by excessive weights;

[0103] formula:

[0104] ;

[0105] The benefit of the formula is that through the adaptive learning rate Adjust weights to reduce overfitting, regularization terms Controlling model complexity to prevent overfitting, The accumulated sum is used to fine-tune the weights and optimize the response speed and accuracy of the network.

[0106] Detailed explanation of the formula and the process of formula calculation and derivation:

[0107] Assume initial , learning rate , the loss function gradient , regularization coefficient , weight decay coefficient , perform the following calculations:

[0108] ;

[0109] ;

[0110] ;

[0111] The result shows that the updated weight is 0.519, which shows the direct impact of weight adjustment on model training, further promoting optimization in the network learning process and ensuring the generalization ability and prediction accuracy of the model.

[0112] S214: Using the updated node parameters and combining the data processing of the network output layer, the impact of environmental parameters on the shelf life of food is calculated, and the predicted value of environmental quality impact is output.

[0113] By updating the node parameters, the system performs complex nonlinear mapping on the data and uses the advanced functions of the output layer, such as activation functions, to convert the processed parameters into specific output values. These output values ​​are transmitted and transformed through multi-layer networks, reflecting the specific impact of changes in environmental parameters on the shelf life of food. The output environmental quality impact prediction value provides a scientific basis for decision-making and demonstrates the complete processing flow from raw data to decision support.

[0114] See also Figure 4 , the specific steps to obtain the optimal neural network configuration result are:

[0115] S311: Based on the predicted value of environmental quality impact, the number of layers, number of neurons and learning rate of each network configuration are extracted through the system interface, and differentiated network architecture combinations are created and integrated into a network architecture combination list;

[0116] Based on the predicted value of environmental quality impact, the system interface extracts various configuration parameters of the neural network, including the number of layers, number of neurons and learning rate. In order to test the combined effect of these parameters, the system first creates a list of network architecture combinations. This list contains a variety of possible configuration combinations. Each combination is based on a specific number of layers, number of neurons and learning rate settings. The purpose of this is to find the most effective network configuration under given environmental quality prediction conditions. By continuously adjusting these parameters, the system can evaluate which configurations are most suitable for processing current environmental data. This process involves a lot of data analysis and parameter adjustment.

[0117] S312: Perform independent network training on each configuration in the network architecture combination list, calculate the performance error of each network configuration using the validation set, evaluate the effectiveness of each network configuration through the performance error, and generate effectiveness evaluation data;

[0118] Through the list of network architecture combinations, each configuration has undergone independent network training. The system uses a set of predetermined verification data to test the performance of these configurations and records the error value of each configuration on the verification set. These error values ​​are then used to evaluate the matching degree of each network configuration. The performance of each network configuration is recorded and analyzed in detail. This analysis process includes the calculation of error values ​​and the quantification of matching degree. Such evaluation helps determine which network configurations can most effectively process input data. Finally, a comprehensive data table containing the performance scores of each configuration is obtained, which shows the potential effectiveness of different network configurations in practical applications.

[0119] S313: Analyze the performance evaluation data and make a comprehensive score based on the performance error and network complexity of each configuration using the formula:

[0120] ;

[0121] Based on the scoring results, the network configuration with the highest matching degree is selected to obtain the optimal configuration result;

[0122] in, Indicates The performance error of each configuration is Represents the complexity of the configuration, is the adjustment coefficient, which is used to balance the influence of performance error and network complexity in the total score. Represents the calculated comprehensive score, which comprehensively considers the performance and complexity of all network configurations and is used to select the optimal neural network configuration. Represents the total number of network configurations involved in the scoring calculation, reflecting the number of evaluated network architecture combinations;

[0123] formula:

[0124] ;

[0125] The benefit of the formula is that it allows us to take into account the error and complexity of each network configuration, where Indicates The performance error of each configuration is Represents the complexity of the configuration, is the adjustment coefficient, which allows us to optimize the choice of the least error while taking into account the complexity of the network, achieving the best balance between performance and complexity.

[0126] Detailed explanation of the formula and the process of formula calculation and derivation:

[0127] Assume that there are three network configurations, with performance errors of 0.05, 0.07, and 0.04, respectively, and complexities of 10, 15, and 20, respectively, and the adjustment coefficient Set it to 0.1. Substituting it into the formula, we get:

[0128] ;

[0129] After calculation, the results are: 6.02, 5.29, 5.72, and the cumulative score is 17.03, indicating that the selected adjustment coefficient effectively balances the error and complexity.

[0130] The results show that the comprehensive score of 17.03 represents the optimal network configuration evaluated based on the given error value and network complexity. The higher the score, the closer the network configuration is to the ideal state, and it can effectively process the input environmental data and ultimately obtain the optimal network configuration result.

[0131] S314: Use the optimal configuration result to adjust the architecture of the neural network to ensure the optimal network configuration and establish the optimal neural network configuration result.

[0132] After obtaining the optimal network configuration results, the system adjusted the overall structure of the neural network to ensure that each part is optimized. This process includes the precise configuration of the number of network layers, the number of neurons and the learning rate. The adjustment process is dynamically optimized according to the predicted values ​​and network performance. Such optimization ensures that the network can maintain optimal performance under various operating conditions. Through precise configuration, the system finally completed the finalization of the network architecture, ensuring that the network can achieve the expected results in practical applications, thereby realizing effective prediction and response to environmental influencing factors.

[0133] See also Figure 5 , the specific steps for obtaining the corrected device control output are:

[0134] S411: Based on the optimal neural network configuration result, input environmental parameters to the neural network structure in real time, predict the impact of food quality based on the input temperature, humidity, and light environmental parameters, and generate food quality prediction results;

[0135] Based on the optimal neural network configuration results, environmental parameters including temperature, humidity, light, etc. are input in real time through the data interface. In the process of predicting the impact of environmental parameters on food quality one by one, first, the system collects real-time data through sensors to monitor parameters such as temperature, humidity, and light in real time. Then, using these data as input, the neural network predicts the possible impact on food quality based on the trained model. This prediction is calculated based on the weights and biases of each layer in the network. The impact of each parameter is converted and output to the next layer through the activation function until the output layer obtains the final food quality prediction result. This process ensures that each input data can be quickly and effectively converted into the corresponding output prediction, ensuring the timeliness and accuracy of the prediction results, which will directly affect subsequent parameter adjustment decisions.

[0136] S412: Using the food quality prediction results, compare the input environmental parameters with the predicted impact, calculate the parameter adjustment amount, readjust the environmental parameters, and calculate the output power and operating time of the equipment using the formula:

[0137] ;

[0138] Obtaining environmental parameter correction values;

[0139] in, Represents the power adjustment factor, which adjusts the power output based on the combined effect of all environmental parameters. Represents the difference between the adjusted temperature parameter and the standard temperature, Represents the difference between the adjusted humidity parameter and the standard humidity, Represents the difference between the adjusted lighting parameters and the standard lighting. Represents the maximum allowable value of temperature, Represents the maximum allowable value of humidity, Represents the weight parameter of the lighting adjustment effect, and adjusts the device power requirement according to the weight of the lighting parameter. Represents the power requirement of the device.

[0140] formula:

[0141] ;

[0142] The benefit of the formula is that it makes adjustments more accurate and efficient by weighted calculation of the impact of different environmental parameters, while taking into account the nonlinear impact of each parameter on the device output, thereby improving the system's responsiveness to environmental changes.

[0143] Detailed explanation of the formula and the process of formula calculation and derivation:

[0144] Assuming the current temperature deviation The maximum permissible temperature deviation is 2°C. For 10°C, humidity deviation The maximum allowable humidity deviation is 5%. 20%, lighting deviation Set the light impact weight to 100Lux. The power adjustment factor is 0.05. is 1.5, plug in the formula:

[0145] ;

[0146] ;

[0147] ;

[0148] ;

[0149] The results show that the output power of the equipment needs to be adjusted to 1.425 times the standard power to adapt to the current environmental conditions, thereby ensuring that the equipment's operating efficiency keeps pace with the environment and minimizing the potential impact of environmental factors on food quality.

[0150] See also Figure 6 , the specific steps for obtaining food storage risk assessment indicators are as follows:

[0151] S511: Based on the corrected device control output, the device operation status and environmental feedback data are obtained in real time, and the parameters of temperature, humidity, and light are collected, combined with the device control status, and integrated to form an environmental monitoring data package;

[0152] Based on the corrected equipment control output, it is key to monitor the adjustment process of a new round of environmental parameters in real time to collect equipment operating status and environmental feedback values. Real-time monitoring of equipment status includes continuous recording of environmental factors such as temperature, humidity, and light. The data acquisition module records the equipment response time and adjustment effect in real time. The monitoring data of environmental parameters is obtained through sensors and transmitted to the central processing unit. The processing unit performs real-time data comparison according to the preset environmental quality standards. By comparing the differences, the system automatically adjusts the corresponding equipment operating parameters, such as adjusting the cooling intensity, humidity control level, etc., to keep the environmental parameters within the optimal range, so as to achieve comprehensive control of the equipment operating status and environmental quality, ensure the stability of the storage environment, and provide the best environmental conditions for food storage.

[0153] S512: Analyze environmental monitoring data packets, use data fusion technology to integrate real-time environmental parameters with equipment output, calculate deviations by comparing preset thresholds with actual values, and evaluate their potential impact on food quality to obtain preliminary risk assessment data;

[0154] In the comprehensive analysis process of environmental monitoring data and equipment control status, real-time environmental parameters and equipment outputs are integrated through advanced data fusion technology, and multivariate statistical analysis methods are used to compare the real-time data of environmental parameters with the preset output standards of the equipment to achieve accurate calculation of deviations. The system compares the real-time data of environmental parameters with historical data, analyzes the changing trends of the data, and evaluates its potential impact on food quality. In this process, the accumulated equipment output and monitoring data provide data support for subsequent risk assessments. Through these analyses, the system can adjust the response strategy in real time, optimize equipment operating efficiency, reduce energy consumption, and improve the level of food safety assurance.

[0155] S513: Analyze the preliminary risk assessment data using the formula:

[0156] ;

[0157] Calculate comprehensive risk indicators, quantify food storage risks, and generate food storage risk assessment indicators;

[0158] in, Represents the actual environmental parameters being monitored, including temperature, humidity, and light. Represents the corresponding safety standard parameters, which are preset environmental parameter standards used to ensure the safety of the food storage environment. is the risk weight of multiple parameters, Represents the food storage risk assessment index, which is a comprehensive risk value. Represents the total number of environmental parameters.

[0159] formula:

[0160] ;

[0161] The benefit of the formula is that it can amplify the impact of environmental parameters on food storage risks by squaring the deviation between each monitored parameter and the standard value and taking a weighted sum, making risk assessment more sensitive and accurate.

[0162] Detailed explanation of the formula and the process of formula calculation and derivation:

[0163] Assume that the environment monitors two parameters: temperature and humidity, where the standard temperature °C, actual temperature °C, standard humidity , actual humidity The weight factors are and .

[0164] Calculate the risk contribution of temperature:

[0165] ;

[0166] Calculate the risk contribution of humidity:

[0167] ;

[0168] The overall risk assessment indicators are:

[0169] ;

[0170] The results show that the risk assessment index of current environmental parameters for food storage is 10.4, indicating that the risk level is high and measures need to be taken to adjust environmental parameters or equipment settings to reduce food storage risks.

[0171] A food storage environment real-time monitoring system, which is used to execute the above-mentioned food storage environment real-time monitoring method, and the system comprises:

[0172] The dynamic adjustment module of environmental parameters obtains the temperature, humidity and light status values ​​based on the real-time monitoring data of the food storage environment, compares the parameters with the standard values, adjusts the fan speed and the duration of the refrigeration equipment, dynamically adjusts the environmental combination, and generates the dynamic adjustment results of environmental parameters;

[0173] The neural network configuration optimization module dynamically adjusts the results of environmental parameters, associates and maps environmental parameters with food shelf life, inputs them into the neural network for weight and bias adjustment, calculates the impact of the environment on food shelf life, and obtains the optimal neural network configuration results;

[0174] The environmental parameter correction and quality prediction module is based on the optimal neural network configuration results, inputs the real-time monitored environmental parameters, calculates the impact on food quality, corrects the environmental parameters, recalculates the equipment output power and duration, and generates environmental parameter correction values;

[0175] The food storage risk assessment module is based on the environmental parameter correction value, monitors the equipment operation status, obtains the environmental parameter feedback value, combines the equipment control status, accumulates the fan and refrigeration equipment output data, and generates the food storage risk assessment index.

[0176] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for real-time monitoring of food storage environment, characterized in that: The following steps are involved: Based on the real-time monitoring data of the food storage environment, the temperature, humidity, and light parameter status values ​​are obtained, and the multi-parameter status is collected and compared with the storage standards. The fan speed and the opening time of the refrigeration equipment are adjusted according to the comparison difference values ​​to generate dynamic adjustment results of environmental parameters; Based on the dynamic adjustment results of the environmental parameters, the environmental parameters and the food shelf life status are mapped, the environmental parameters are input into the neural network connection structure, the weights and deviations are adjusted step by step, the node parameters are updated, and the environmental quality impact prediction value is obtained; Based on the predicted value of environmental quality impact, the number of layers, number of neurons and learning rate of the neural network are cross-combined, the matching degree of each combination is scored, the combination structure with the highest matching degree is selected, and the optimal neural network configuration result is established; Based on the optimal neural network configuration result, input environmental parameters to the neural network in real time, predict the impact of environmental parameters on food quality one by one, readjust the input parameters, use the adjusted parameters to calculate the output power and duration of the equipment, and generate environmental parameter correction values; Based on the environmental parameter correction value, the equipment operating status and environmental feedback value are called up to monitor a new round of environmental parameters, combine temperature, humidity, and light with the equipment control status, accumulate all equipment output and monitoring data, and generate food storage risk assessment indicators.

2. The method for real-time monitoring of food storage environment according to claim 1, characterized in that: The dynamic adjustment results of the environmental parameters include the fan operating speed, the refrigeration equipment startup time, and the environmental parameter difference value; the environmental quality impact prediction value includes the temperature prediction value, the humidity prediction value, and the light prediction value; the optimal neural network configuration result includes the number of neural network layers, the number of neurons, and the learning rate; the environmental parameter correction values ​​include the equipment output power and the equipment output time; the food storage risk assessment indicators include temperature risk, humidity risk, light risk, and equipment operating status.

3. The method for real-time monitoring of food storage environment according to claim 2, characterized in that: The steps for obtaining the result of the dynamic adjustment of the environmental parameters are specifically as follows: Collect real-time monitoring data, including temperature, humidity, and light parameters, obtain the current state value of the food storage environment through the sensor network, and generate real-time environmental status records; Compare the temperature, humidity, and light conditions in the real-time environmental status record with the preset storage standard, calculate the difference values ​​of multiple parameters, and obtain the environmental difference value result; According to the environmental difference value results, the fan speed and the working time of the refrigeration equipment are adjusted using the formula: ； Calculate the adjustment amount of fans and refrigeration equipment, and generate dynamic adjustment results of environmental parameters ; in, represents the temperature difference value, Represents the humidity difference value, Represents the illumination difference value, is the standard threshold value of temperature, is the standard threshold value of humidity, is the standard threshold of illumination, is the temperature regulation coefficient, is the humidity adjustment factor, is the light regulation coefficient.

4. The method for real-time monitoring of food storage environment according to claim 3, characterized in that: The steps for obtaining the environmental quality impact prediction value are specifically as follows: Based on the dynamic adjustment result of the environmental parameters, key temperature, humidity and light parameters are extracted, and the parameters are input into the input layer of the neural network through the data interface to generate an input layer data carrier; The data is transmitted layer by layer to each hidden layer of the neural network through the input layer data carrier, and the weights and biases of the nodes of each layer are adjusted in real time. The adjusted weights and biases of each layer are used to generate intermediate node parameters; Based on the intermediate node parameters, the formula is adopted: ； Adjust weights and biases, update the weight of each node, and obtain updated node parameters; in, Represents the weight before adjustment, which is the main parameter for transmitting information in the neural network and directly affects the output of the network. Represents the adjusted weight, which is the new weight value obtained by adjusting the algorithm update to optimize the prediction or classification accuracy of the network. Represents the learning rate, controls the step size of weight adjustment, and directly affects the speed and stability of network training. It is the derivative of the loss function with respect to the weight, indicating the sensitivity of the loss function to the change under the current weight, and is used to guide the optimization direction of the weight. is a regularization parameter used for weight attenuation, preventing network overfitting and enhancing the generalization ability of the model. is the weight decay coefficient, which is used to adjust the influence of each weight and reduce the problem of network dependence on specific paths caused by excessive weights; The updated node parameters are used in combination with data processing of the network output layer to calculate the impact of environmental parameters on the shelf life of food, and output a predicted value of environmental quality impact.

5. The method for real-time monitoring of food storage environment according to claim 4, characterized in that: The steps for obtaining the optimal neural network configuration result are specifically as follows: Based on the environmental quality impact prediction value, the number of layers, number of neurons and learning rate of each network configuration are extracted through the system interface, and differentiated network architecture combinations are created and integrated into a network architecture combination list; Performing independent network training on each configuration in the network architecture combination list, calculating the performance error of each network configuration using a validation set, evaluating the effectiveness of each network configuration through the performance error, and generating effectiveness evaluation data; The performance evaluation data is analyzed and a comprehensive score is given based on the performance error and network complexity of each configuration using the formula: ； Based on the scoring results, the network configuration with the highest matching degree is selected to obtain the optimal configuration result; in, Indicates The performance error of each configuration is Represents the complexity of the configuration, is the adjustment coefficient, which is used to balance the influence of performance error and network complexity in the total score. Represents the calculated comprehensive score, which comprehensively considers the performance and complexity of all network configurations and is used to select the optimal neural network configuration. Represents the total number of network configurations involved in the scoring calculation, reflecting the number of evaluated network architecture combinations; Using the optimal configuration result, the neural network architecture is adjusted to ensure the optimization of the network configuration and establish the optimal neural network configuration result.

6. The method for real-time monitoring of food storage environment according to claim 5, characterized in that: The specific steps for obtaining the corrected device control output are as follows: Based on the optimal neural network configuration result, input environmental parameters to the neural network structure in real time, predict the impact of food quality based on the input temperature, humidity, and light environmental parameters, and generate food quality prediction results; Using the food quality prediction results, the relationship between the input environmental parameters and the predicted impact is compared, the parameter adjustment amount is calculated, the environmental parameters are readjusted, and the output power and operating time of the equipment are calculated using the formula: ； Obtaining environmental parameter correction values; in, Represents the power adjustment factor, which adjusts the power output based on the combined effect of all environmental parameters. Represents the difference between the adjusted temperature parameter and the standard temperature, Represents the difference between the adjusted humidity parameter and the standard humidity, Represents the difference between the adjusted lighting parameters and the standard lighting. Represents the maximum allowable value of temperature, Represents the maximum allowable value of humidity, Represents the weight parameter of the lighting adjustment effect, and adjusts the device power requirement according to the weight of the lighting parameter. Represents the power requirement of the device.

7. The method for real-time monitoring of food storage environment according to claim 6, characterized in that: The steps for obtaining the food storage risk assessment index are specifically as follows: Based on the corrected device control output, the device operation status and environmental feedback data are obtained in real time, and the parameters of temperature, humidity, and light are collected, combined with the device control status, to form an environmental monitoring data package; Analyze the environmental monitoring data packets, use data fusion technology to integrate real-time environmental parameters with equipment outputs, calculate deviations by comparing preset thresholds with actual values, and evaluate their potential impact on food quality to obtain preliminary risk assessment data; The preliminary risk assessment data were analyzed using the formula: ； Calculate comprehensive risk indicators, quantify food storage risks, and generate food storage risk assessment indicators; in, Represents the actual environmental parameters being monitored, including temperature, humidity, and light. Represents the corresponding safety standard parameters, which are preset environmental parameter standards used to ensure the safety of the food storage environment. is the risk weight of multiple parameters, Represents the food storage risk assessment index, which is a comprehensive risk value. Represents the total number of environmental parameters.

8. A food storage environment real-time monitoring system, characterized in that: According to the method for real-time monitoring of a food storage environment according to any one of claims 1 to 7, the system comprises: The environmental parameter dynamic adjustment module obtains the temperature, humidity and light status values ​​based on the real-time monitoring data of the food storage environment, compares the parameters with the standard values, adjusts the fan speed and refrigeration equipment duration, dynamically adjusts the environment combination, and generates the dynamic adjustment results of the environmental parameters; The neural network configuration optimization module dynamically adjusts the environmental parameters based on the environmental parameters, associates and maps the environmental parameters with the food shelf life status, inputs the environmental parameters into the neural network for weight and bias adjustment, calculates the impact of the environment on the food shelf life, and obtains the optimal neural network configuration result; The environmental parameter correction and quality prediction module inputs the real-time monitored environmental parameters based on the optimal neural network configuration result, calculates the impact on food quality, corrects the environmental parameters, recalculates the equipment output power and duration, and generates environmental parameter correction values; The food storage risk assessment module monitors the equipment operation status, obtains the environmental parameter feedback value, and generates food storage risk assessment indicators based on the environmental parameter correction value, the accumulated fan and refrigeration equipment output data combined with the equipment control status.

Images